How to Do a Competitive Analysis (Step-by-Step)

"Analyze our competitors" is a topic. "Decide whether to launch a free tier in Q3" is a decision. The second framing tells you exactly which competitors to study (those with free tiers), which metrics matter (conversion from free to paid, activation rate signals, content volume aimed at free users), and which can be ignored.

Write the decision as a single sentence at the top of your working doc. Every chart, screenshot, and note that follows should connect back to it. If a finding does not, park it in an appendix or delete it.

Three scope dimensions matter most. First, time window: 90 days is enough for most tactical questions, 12 months for strategic ones. Second, channels: pick two or three (for example, Instagram, LinkedIn, and competitor blogs) rather than chasing every platform. Third, geography: a global view often hides the only market you actually sell in.

Document the scope in writing and share it with stakeholders before you start. This protects you from the classic mid-project request: "Can you also add TikTok? And Germany? And our top five resellers?" Scope creep is the number-one reason competitive analyses ship late and shallow.

Decide now whether the deliverable is a 6-slide deck, a Notion doc, a Loom walkthrough, or a live workshop. The format changes how you collect data. A workshop needs printable matrices. A deck needs charts. A Loom needs a script. Choosing at the end forces rework.

With the goal, scope, and format locked, you can move to the part most teams rush: figuring out who your competitors actually are.

Direct competitors sell a similar product to a similar buyer. If you run a mid-market project management tool, Asana and ClickUp are direct. Pull these from three sources: closed-lost deal notes (the most reliable), G2 or Capterra category pages, and branded search comparison queries ("yourbrand vs ___").

Most teams should pick three to five direct competitors. More than seven and you cannot maintain depth. Fewer than three and you miss patterns.

Indirect competitors solve the same problem with a different shape. For that project management tool, indirect competitors include Notion, Google Sheets, and "a Slack channel and good intentions". These steal budget and attention even though they would never appear in a head-to-head G2 comparison.

Indirect competitors matter most for content and positioning. They tell you which behaviors and substitutes you must displace, not just which logos you must out-feature.

Start with what is freely accessible: competitor websites, pricing pages, changelogs, careers pages, customer case studies, podcasts, earnings calls (for public companies), and review sites like G2, Capterra, and Trustpilot. These give you positioning, product roadmap signals, and customer language in their own words.

Set up a structured doc per competitor with consistent fields: tagline, target ICP, pricing tiers, headline features, three recent customer wins, and three recent critical reviews. Consistency is what makes the data comparable later.

Social channels are where competitor strategy plays out in public. Posting cadence, content themes, engagement patterns, ad creative, and influencer partnerships all show intent before any press release does. Facebook, Instagram, and Twitter/X each reveal different things: Facebook for paid creative (the Ads Library is a goldmine), Instagram for brand and product launches, Twitter/X for narrative and announcements.

Manual tracking is fine for one or two competitors over a short window. Past that, it breaks. Pulling 90 days of posts across five competitors on three platforms means several thousand data points, screenshot drift, and inconsistent metrics. This is where automated tools earn their cost.

For SEO and ad data, Ahrefs, Semrush, and Similarweb are the well-known options. For social, the picture is more fragmented. Native tools (Facebook Insights, Instagram Insights, Twitter Analytics) only show you your own data, not competitors.

This is the gap Competitor Analyzer was built to fill: automatic daily tracking of competitor posts, engagement, and landing-page changes across Facebook, Instagram, and Twitter/X, with cross-platform comparison in one view. The point is not to replace judgment, but to remove the four hours per week of copy-paste work that drains analyst attention from actual analysis. Whichever tools you pick, write down your stack and refresh schedule. Future-you will thank present-you when the next analysis takes half the time.

A solid profile has eight fields: company snapshot (HQ, headcount, funding), positioning statement (in their own words), ICP and segments, pricing model and tiers, top three product features they push, content and channel mix, recent strategic moves (last 90 days), and a short "so what" paragraph from you.

The "so what" paragraph is the most important and the most skipped. It forces you to answer: based on this profile, what should we do, watch for, or stop doing? Without it, the profile is just a Wikipedia entry.

Competitor profiles decay. Pricing changes. New features ship. Founders leave. A profile from twelve months ago is closer to fiction than fact. Set a refresh cadence: monthly for top three direct competitors, quarterly for the rest.

Tools that auto-track landing-page changes and post activity make this maintenance almost free. With profiles built and current, you can do the analytical work most teams skip: structured SWOT.

Strengths and weaknesses are about the competitor's current state. Strengths might include a defensible feature, a pricing edge, a beloved community, or a distribution channel you cannot easily access. Weaknesses might include a confusing onboarding flow, poor mobile experience, a product gap, or a wave of recent 1-star reviews about support.

Pull weaknesses from review sites, Reddit threads, and complaint patterns on Twitter/X. Customers will tell you exactly what is broken if you read their words instead of the competitor's marketing.

Opportunities and threats are about the environment around the competitor. Opportunities for them might include a new buyer segment opening up, a regulatory tailwind, or an adjacent category they can extend into. Threats might include platform algorithm changes (Instagram's reach drops, iOS privacy shifts), new entrants, or a shift in buyer preferences.

Be careful not to mix internal and external. "Their content team is small" is a weakness. "Content marketing as a channel is getting more crowded" is a threat. The distinction matters because the response is different.

Every SWOT entry should produce at least one candidate action for your team. A competitor weakness is a positioning angle for you. A competitor strength is something to either neutralize or route around. A threat to them might be a tailwind for you.

End each SWOT with a five-line summary: top three actions for us, top two things to monitor. That summary is what survives into the final report.

The classic 2x2 plots competitors on two axes: for example, price (low to high) versus product depth (light to heavy), or self-serve versus sales-led, or horizontal versus vertical. Pick axes that match the strategic question. If you are deciding on pricing, one axis must be price.

Plot each competitor based on evidence, not gut. A 2x2 with eight logos clustered in one corner is not a failed map; it is the answer. It says the market has a gap, and that gap is your opportunity.

For tactical decisions, a feature matrix beats a 2x2. Rows are features, columns are competitors, cells are checkmarks, partials, or notes. Same for pricing: rows are tiers and limits, columns are competitors. These look boring and are extremely useful.

Keep them short. A matrix with 60 rows is unreadable. Pick the 10 to 15 features or pricing dimensions that actually drive buying decisions. The rest is noise.

On the marketing side, two maps matter most. First, share of voice: what percentage of category conversation, mentions, or branded search each competitor owns. Second, content map: which topics, formats, and channels each competitor invests in.

Both are easier with cross-platform tracking. Pulling 90 days of posts across Facebook, Instagram, and Twitter/X manually for five competitors is a week of work. Automated tools (Competitor Analyzer included) collapse that to a dashboard, which means you spend the week on interpretation instead of collection.

Look for gaps in four places. Product gaps: features customers ask for that no competitor ships well. Segment gaps: buyer types nobody targets directly (often mid-market, often a specific industry vertical). Channel gaps: distribution channels competitors ignore (a podcast, a Reddit community, an industry event). Message gaps: things customers care about that nobody is talking about.

Message gaps are the easiest to act on and the most underrated. If five competitors all market on "speed" and customer reviews keep mentioning "trust", trust is a positioning lane sitting empty.

Not every gap is an opportunity for you. Score each candidate on three criteria: impact (how much does this move the business?), fit (do we have the team, product, and brand to win here?), and effort (how many quarters and people will it take?). A simple 1-5 score on each, multiplied, gives you a ranked list.

Be brutal about fit. The most expensive mistake in competitive analysis is chasing a gap that is real but not yours. If your team is six people and the gap requires a 30-person enterprise sales motion, the gap belongs to someone else.

End this step with three to five opportunities, ranked, with a one-line rationale each. Thirty opportunities is a wish list. Three is a strategy. The discipline of cutting is the work.

With opportunities chosen, the last job is packaging the analysis so it actually drives action.

Start with a one-page TL;DR: the decision question, your three to five recommendations, and the single most important chart. Anyone who reads only this page should still walk away knowing what you think and why.

Then layer detail in the order: landscape map, top opportunities, competitor profiles, methodology and data sources. Methodology goes last, not first. Stakeholders trust the work because of the recommendations, not because of the methodology slide.

"Competitor X posts more on Instagram than we do" is an observation. "We should increase Instagram posting from 2 to 4 times per week and shift 60% of new content to Reels, based on the engagement gap shown in slide 8" is a recommendation. Stakeholders need the second.

Yes, you might be wrong. Recommend anyway. A wrong, specific recommendation can be debated and corrected. A vague observation cannot. Vagueness is what gets reports shelved.

A 70%-complete analysis delivered in week three beats a 95%-complete analysis delivered in week eight. Markets move. Competitor pricing shifts. Your CEO's question changes. Ship a v1, get reactions, then deepen the parts that matter.

Set a standing refresh cadence: a quarterly update of the full analysis, monthly snapshots of social and content moves. Features like AI-powered alerts and a weekly market summary keep that cadence cheap to maintain. With the report shipped and the cadence set, the only thing left is avoiding the traps that sink most analyses.

Teams build a 200-row feature matrix and conclude they need 23 new features to catch up. This almost always misreads the market. Customers rarely buy on feature parity; they buy on positioning, trust, and one or two killer capabilities.

Fix: cap your feature matrix at 15 rows, and force a separate "jobs to be done" view alongside it. The matrix shows what competitors built. The jobs view shows what customers actually want done.

An analysis based on data from six months ago is worse than no analysis. It is wrong with confidence. Pricing has changed, leadership has rotated, the product roadmap has shifted.

Fix: timestamp every data point in your working doc. Set automated tracking on the things that change most (pricing pages, social posts, ad creative, landing pages). A tool like Competitor Analyzer that flags landing-page changes daily means you find out about a competitor's repositioning the day it ships, not the quarter after.

The report is delivered. Everyone nods. Three months later, nothing has changed. This is the most common failure mode and the saddest, because the work was good.

Fix: every recommendation must have a single named owner and a deadline before the deck is presented. "Marketing should consider" is not an owner. "Priya, by June 15" is.

Analysts often start with a hypothesis ("Competitor X is beating us on content") and unconsciously gather data that supports it. The findings then justify a decision the team already wanted to make.

Fix: write down two competing hypotheses at the start, and assign yourself to disprove the one you believe. If the data still supports it after honest attack, your conviction is earned. If not, you saved your company a wrong bet, which is exactly what a competitive analysis is supposed to do.